Courtney Murray!

####Run through this example and try to understand what is going on with the data

####So, lets load RMarkdown

library(rmarkdown)

set messages to FALSE on everything (prevents certain boring things from being shown in the results)

knitr::opts_chunk$set(echo = TRUE, message=FALSE,warning=FALSE,collapse = TRUE)

PACKAGES

library(reshape2)
library(ggplot2)
library(dplyr)
library(plotly)
library(viridis)
library(data.table)
library(pheatmap)
library(tidyverse)
library(ggthemes)
library(clipr)
library(tidyr)
library(Rcpp)


# load the dataset
breast_cancer_cells <- read.csv("breast_cancer_cells.csv")

COLORS

mycolors<-c(viridis(15))
felix_cols<-mycolors[c(5,2)]
felix_4cols<-mycolors[c(15,10,8,2)]
plain_cols1<-c("blue","gray")
plain_cols2<-c("red","gray")

pats_cols<-colorRampPalette(c("#FDE725FF", "white","#440154FF"))(21)
leos_cols<-colorRampPalette(c("white","blue"))(10)

LOAD DATA AND MAKE AN OVERALL HEATMAP

## load the dataset 
C19_data<-read_csv(file="breast_cancer_cells.csv")

#and then view it 
view(C19_data)

now let’s visualize the dataset and look for initial trends. We can do this by making a matrix so and then a heatmap to visualize the data

#make a matrix of just raw the data

breast_mat1 <- breast_cancer_cells %>%
  dplyr::select(
    MCF10A_1, MCF10A_2,
    MCF7_1, MCF7_2,
    MDA231_1, MDA231_2,
    MDA468_1, MDA468_2,
    SKBR3_1, SKBR3_2
  ) %>%
  as.matrix() %>%
  round(2)

# what does that look like?

head(breast_cancer_cells)
##   Gene_Symbol
## 1         NES
## 2      ZC3HC1
## 3   CAMSAP1L1
## 4      COBRA1
## 5       HAUS6
## 6      CHCHD4
##                                                                                                Description
## 1                                                                  Nestin OS=Homo sapiens GN=NES PE=1 SV=2
## 2                                Isoform 2 of Nuclear-interacting partner of ALK OS=Homo sapiens GN=ZC3HC1
## 3             Isoform 2 of Calmodulin-regulated spectrin-associated protein 2 OS=Homo sapiens GN=CAMSAP1L1
## 4                                         Negative elongation factor B OS=Homo sapiens GN=COBRA1 PE=1 SV=1
## 5                                 Isoform 3 of HAUS augmin-like complex subunit 6 OS=Homo sapiens GN=HAUS6
## 6  Isoform 2 of Mitochondrial intermembrane space import and assembly protein 40 OS=Homo sapiens GN=CHCHD4
##   Peptides MCF10A_1 MCF10A_2 MCF7_1 MCF7_2 MDA231_1 MDA231_2 MDA468_1 MDA468_2
## 1        7     9.54     4.58   5.07   5.42    25.43    27.42     4.56     3.88
## 2        2    14.00    11.58   6.49   6.64     9.80    10.31     6.84     8.75
## 3        5    10.22     8.29  11.55  10.82    12.48    10.11     9.54     8.46
## 4        3     9.00     6.35  13.40  14.82    14.94    11.33     6.87     7.92
## 5        5     8.21     4.44  12.08   9.82    16.51    12.34    15.43     8.50
## 6        5    12.51    15.84   8.05   8.38     6.78     8.46     4.48     7.18
##   SKBR3_1 SKBR3_2 pvalue_MCF7_vs_MCF10A pvalue_MDA231_vs_MCF10A
## 1    8.46    5.64            0.54152658              0.01849512
## 2   11.91   13.67            0.03589005              0.15732968
## 3    9.10    9.43            0.20243066              0.31440425
## 4    8.54    6.82            0.05045427              0.13484982
## 5    7.93    4.75            0.16972788              0.10216410
## 6   13.27   15.05            0.07037866              0.07222737
##   pvalue_MDA468_vs_MCF10A pvalue_SKBR3_vs_MCF10A
## 1              0.37423781              0.9983818
## 2              0.08358401              0.9980974
## 3              0.83825377              0.9967306
## 4              0.86215724              0.9958838
## 5              0.28867772              0.9957949
## 6              0.06005626              0.9948536

## make a heatmap from the data
pheatmap(breast_mat1,
         color = pats_cols,
         cellwidth = 30,
         cellheight = 0.03,
         cluster_cols = FALSE,
         cluster_rows = TRUE,
         legend = TRUE,
         fontsize = 7,
         scale = "column")

main = "Heatmap of Protein Expression Across Breast Cancer Cell Lines"
##INTERPRETATION##
## What can you see in this figure? are the repeated measures/reps similar or different? What does this say about the precision and accuracy of them? 
##How does the control compare to the variables? Is this what you might expect? Why? What would you look for in the literature to support this idea?  

#looks like _______ line  is different than the other cells lines.  We will take a closer look at that in the next section, once we've done some further data manipulations

Data Manipulation

## first, make new columns that combine the two reps for each variable by averaging them  
breast_data2 <- breast_cancer_cells %>% mutate(
  mean_MCF10A = ((MCF10A_1 + MCF10A_2) / 2),
  mean_MCF7   = ((MCF7_1 + MCF7_2) / 2),
  mean_MDA231 = ((MDA231_1 + MDA231_2) / 2),
  mean_MDA468 = ((MDA468_1 + MDA468_2) / 2),
  mean_SKBR3  = ((SKBR3_1 + SKBR3_2) / 2))


 
view(breast_data2)
#did it create your new columns?

## then make some new columns that store the log ratios of the variable means you just created, as compared to the control

breast_data2 <- breast_data2 %>% mutate(
  log_MCF7   = log2(mean_MCF7 / mean_MCF10A),
  log_MDA231 = log2(mean_MDA231 / mean_MCF10A),
  log_MDA468 = log2(mean_MDA468 / mean_MCF10A),
  log_SKBR3  = log2(mean_SKBR3 / mean_MCF10A))
 
colnames(breast_data2)
##  [1] "Gene_Symbol"             "Description"            
##  [3] "Peptides"                "MCF10A_1"               
##  [5] "MCF10A_2"                "MCF7_1"                 
##  [7] "MCF7_2"                  "MDA231_1"               
##  [9] "MDA231_2"                "MDA468_1"               
## [11] "MDA468_2"                "SKBR3_1"                
## [13] "SKBR3_2"                 "pvalue_MCF7_vs_MCF10A"  
## [15] "pvalue_MDA231_vs_MCF10A" "pvalue_MDA468_vs_MCF10A"
## [17] "pvalue_SKBR3_vs_MCF10A"  "mean_MCF10A"            
## [19] "mean_MCF7"               "mean_MDA231"            
## [21] "mean_MDA468"             "mean_SKBR3"             
## [23] "log_MCF7"                "log_MDA231"             
## [25] "log_MDA468"              "log_SKBR3"
#did it create your new columns?

## ALTERNATIVE CHOICE: make a matrix of just log values the data and a heat map of that. How do the two heat maps compare? This is not possible if your values contain 0's

breast_mat2 <- breast_data2 %>%
  dplyr::select(log_MCF7:log_SKBR3) %>%
  as.matrix() %>%
  round(2)

pheatmap(
  breast_mat2,
  color = pats_cols,
  cellwidth = 30,
  cellheight = 0.03,
  cluster_cols = FALSE,
  cluster_rows = TRUE,
  legend = TRUE,
  fontsize = 7,
  scale = "column")

Diving deeping with VOLCANO PLOTS

##BASIC VOLCANO PLOT

## Add a column that stores the negative log of the pvalue of interest (in this case, _____)
breast_data2 <- breast_data2 %>% mutate(
 neglog_MCF7 = -log10(pvalue_MCF7_vs_MCF10A)
)

## Use ggplot to plot the log ratio of ____ against the -log p-value ____
volcano_plot <- breast_data2 %>%
  ggplot(aes(
    x = log_MCF7,
    y = neglog_MCF7,
    text = Gene_Symbol
  )) +
  geom_point(alpha = 0.7, color = "royalblue")

#to view it, type: 
volcano_plot

  
## BETTER VOLCANO PLOT

## Define the significant ones (by using ifelse and setting # parameters) so they can be colored
breast_data2 <- breast_data2 %>% mutate(
  significance = ifelse(
    (log_MCF7 > 1 & neglog_MCF7 > -log10(0.05)), "UP",
    ifelse(
      (log_MCF7 < -1 & neglog_MCF7 > -log10(0.05)), "DOWN",
      "NOT SIG")))

## Some standard colors

cols<-c("red","gray","blue")
## volcano plot as before with some added things


better_volcano_plot <- breast_data2 %>%
  ggplot(aes(
    x = log_MCF7,
    y = neglog_MCF7,
    text = Gene_Symbol,
    color = significance
  )) +
  geom_point(alpha = 0.7) +
  scale_color_manual(values = cols) +
  xlim(-6, 6) +
  theme_bw() +
  theme(axis.text = element_text(colour = "black", size = 14)) +
  theme(text = element_text(size = 14)) +
  labs(
    x = "log2 fold change (MCF7 vs MCF10A)",
    y = "-log10(p-value)")

  
#to view it, type
better_volcano_plot

#check viewer and / or plots to see it 

## use ggplotly to see hover over the points to see the gene names. Record these for the next step 

ggplotly(better_volcano_plot)

##INTERPRETATION##
## Significant up regulated are: ___, ___ and _____
## Significant down regulated are: _____. _____ and ____
#Why? How?

Barplots of significant points of interest

EXAMPLES OF A COUPLE PROTEINS or GENES

# Make a pivot longer table of the C19 data for Control_2h_1:Virus_24h_2

breast_long <- breast_data2 %>%
  pivot_longer(
    cols = c(
      MCF10A_1, MCF10A_2,
      MCF7_1, MCF7_2,
      MDA231_1, MDA231_2,
      MDA468_1, MDA468_2,
      SKBR3_1, SKBR3_2
    ),
    names_to = "variable",
    values_to = "value"
  )

Examples_Down <- breast_long %>%
  filter(Gene_Symbol %in% c("CTSZ", "APOA1", "HLA-A"))

Example_plot_down <- Examples_Down %>%
  ggplot(aes(
    x = factor(
      variable,
      levels = c(
        "MCF10A_1", "MCF10A_2",
        "MCF7_1", "MCF7_2",
        "MDA231_1", "MDA231_2",
        "MDA468_1", "MDA468_2",
        "SKBR3_1", "SKBR3_2"
      )
    ),
    y = value
  )) +
  geom_bar(stat = "identity", fill = "red") +
  facet_wrap(~Gene_Symbol) +
  theme_bw() +
  theme(axis.text = element_text(colour = "black", size = 10)) +
  theme(text = element_text(size = 14)) +
  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
  labs(x = "sample", y = "relative intensity")

Example_plot_down

#check viewer and / or plots to see it 

##INTERPRETATION## ## What can you see in this figure? are the repeated measures/reps similar or different? What does this say about the precision and accuracy of them? ##How does the control compare to the variables? Is this what you might expect? Why? What would you look for in the literature to support this idea?

## Same process for the upregulated ones

Examples_Up <- breast_long %>%
  filter(Gene_Symbol %in% c("UCHL1", "C17orf28", "AGR3"))
Example_plot_up <- Examples_Up %>%
  ggplot(aes(
    x = factor(
      variable,
      levels = c(
        "MCF10A_1", "MCF10A_2",
        "MCF7_1", "MCF7_2",
        "MDA231_1", "MDA231_2",
        "MDA468_1", "MDA468_2",
        "SKBR3_1", "SKBR3_2"
      )
    ),
    y = value
  )) +
  geom_bar(stat = "identity", fill = "blue") +
  facet_wrap(~Gene_Symbol) +
  theme_bw() +
  theme(axis.text = element_text(colour = "black", size = 10)) +
  theme(text = element_text(size = 14)) +
  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
  labs(x = "sample", y = "relative intensity")

Example_plot_up

#check viewer and / or plots to see it 

##INTERPRETATION## ## What can you see in this figure? are the repeated measures/reps similar or different? What does this say about the precision and accuracy of them? ##How does the control compare to the variables? Is this what you might expect? Why? What would you look for in the literature to support this idea?

#interpretation HINT:insert a chunk and create two seprate lines of code that filter for your specific upregulated genes/proteins of interest and selects for only their gene symbols and descriptions. Do this for the downregulated as well. This will generate two list of the descriptors for each gene of interest, helping you understand your figures. Be sure to view it, not just ask for the head of the table generated.

WRAP UP

##now you can knit this and publish to save and share your code. Use this to work with either the brain or breast cells and the Part_C_template to complete your lab 6 ELN. ##Annotate when you have trouble and reference which line of code you need help on ## good luck and have fun!